hynix666

Adversarial — AI skill for Claude Code

AI community

Adversarial design critique of code or a diff, via the local Ollama model.

How to install Adversarial

Installs to ~/.claude/skills/hynix666-ollama-reviewer-adversarial/SKILL.md

Terminal

mkdir -p ~/.claude/skills/hynix666-ollama-reviewer-adversarial && curl -fsSL https://raw.githubusercontent.com/hynix666/ollama-reviewer/HEAD/commands/adversarial.md -o ~/.claude/skills/hynix666-ollama-reviewer-adversarial/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.

What Adversarial does

description: Adversarial design critique of code or a diff, via the local Ollama model

Run an adversarial review — the model is told to assume the design is wrong and to attack its assumptions rather than its syntax.

python ~/.claude/skills/ollama-reviewer/scripts/cli.py review --adversarial
python ~/.claude/skills/ollama-reviewer/scripts/cli.py review --adversarial --file $ARGUMENTS

Use the `--file` form when $ARGUMENTS names paths; otherwise review the current diff. Pass

Alternatives in AI

Codex — openai-codex Adversarial code review, Codex CLI integration, cross-model analysis 9.6k ★
Chat Ollama — ChatOllama is an open source agentic app for running AI agents across local and hosted models 3.5k ★
DontFeedTheAI — Transparent anonymization proxy for AI-assisted pentesting 654 ★